Nobuyuki Nakamori

Image feature analysis and computer‐aided diagnosis in digital radiography: Effect of digital parameters on the accuracy of computerized analysis of interstitial disease in digital chest radiographs

We are developing a computerized method for measurement of lung texture in digital chest radiographs for detection and characterization of interstitial disease. Physical texture measures are obtained from analysis of the power spectrum of the lung texture. We have investigated the effect of digital parameters such as pixel size, regions of interest size, the number of quantitation levels, and the peak frequency of the visual system response, as well as the effect of the unsharp masking technique on the performance of this computerized method. We calculated the texture measures by changing digital parameters for 100 normal lungs and 100 abnormal lungs in our database. Receiver operating characteristic (ROC) curves were employed for evaluation of the performance of this computerized method for distinguishing between normal and abnormal lungs. We used the area under the ROC curve to compare the detection accuracy for interstitial infiltrates. We believe that the results of this study may be useful as a guide in the design of computerized schemes for lung texture analysis in digital chest radiographs.

Computer-aided diagnosis in chest radiology

Digital radiography offers several important advantages over conventional systems, including abilities for image manipulation, transmission, and storage. In the long term, however, the unique ability to apply artificial intelligence techniques for automated detection and quantitation of disease may have an even greater impact on radiologic practice. Although CAD is still in its infancy, the results of several recent studies clearly indicate a major potential for the future. The concept of using computers to analyze medical images is not new, but recent advances in computer technology together with progress in implementing practical digital radiography systems have stimulated research efforts in this exciting field. Several facets of CAD are presently being developed at the University of Chicago and elsewhere for application in chest radiology as well as in mammography and vascular imaging. To date, investigators have focused on a limited number of subjects that have been, by their nature, particularly suitable for computer analysis. There is no aspect of radiologic diagnosis that could not potentially benefit from this approach, however. The ultimate goal of these endeavors is to provide a system for comprehensive automated image analysis, the results of which could be accepted or modified at the discretion of the radiologist.

Effect of heart-size parameters computed from digital chest radiographs on detection of cardiomegaly. Potential usefulness for computer-aided diagnosis.

Heart size is an important and useful diagnostic parameter in chest radiographs. However, there is a large variation in the subjective judgment of cardiac enlargement (cardiomegaly). To reduce this subjective element, the authors are developing an automated system for quantitative analysis of heart size in digital chest radiographs. Four hundred chest radiographs were reviewed initially by two radiologists and were classified into two groups: those with and those without cardiomegaly. Another radiologist reviewed 47 images which were not classified consistently in the initial review. The authors used these radiographs to construct a database for determination of cardiomegaly. Numerous parameters related to heart size were obtained in a semiautomated analysis of these radiographs, and the use of each parameter for detection of cardiomegaly was evaluated by means of receiver operating characteristic analysis. The authors also examined whether the accuracy would be improved when they applied multivariate analysis to a pair of parameters. From the analyses of the individual parameters, the automatically determined cardiothoracic ratio was found to be the single most effective measure for detecting cardiomegaly in chest radiographs. However, multivariate analysis provided results superior to those with an individual parameter.

Utilization Of Digital Image Data For Computer-aided Diagnosis

A number of computerized schemes being developed for computer-aided diagnosis (CAD) in our laboratory are reviewed. In distinguishing between normal and abnormal lungs with interstitial infiltrates in chest images, the computerized classification method provided the ROC curve that is comparable to or superior to that obtained by an average radiologist. Our computerized detection schemes indicated truepositive detection rate of approximately 70% for subtle lung nodules in chest radiographs and 90% for subtle clustered microcalcifications in mammograms, although several false positives were detected in each image. The automatically computed outlines of the heart shadows in chest radiographs were very similar to the contours traced by radiologists, and were used to obtain parameters related to the size and area of the projected heart. By using an iterative deconvolution technique, opacified vessels larger than 0.5 mm in DSA images were measured with an accuracy of approximately 0.1mm. The vascular structures in angiograms were tracked accurately and automatically by using a double-square-box region-of-search method.

Automated Computerized Analysis Of Radiographic Images: An Intelligent Component Of IMAC To Aid Diagnosis

A weakness of the impact of the current IMAC and PACS on diagnostic radiology is partly due to the lack of an “intelligent” component which could assist radiologists in their diagnosis. An intelligent component of PACS can be implemented by automated computerized analysis of radiographic images. Computer output from quantitative analysis of digital images can be used as a “second opinion” to alert the radiologist by indicating potential lesion sites and/or by providing objective measurements of normal and abnormal patterns. The use of computer output in this manner is expected to improve diagnostic accuracy by reducing false negatives and U) improve the overall reproducibility of image interpretation. Current results of computerized automated analysis are demonstrated for the identification of lung nodules and pneumothoraces, and for the assessment of interstitial disease and cardiomegaly in chest radiography. These early results are very encouraging. Computer-aided diagnosis, which refers to a diagnosis made by a radiologist who utilizes the computer output of quantitative image analysis, may be a useful component of PACS which can provide practical benefits for radiologists.